Nucleic Acids Research, Vol. 18, No. 19 5677

k.) 1990 Oxford University Press

The promoter of the human parathyroid hormone gene contains a functional cyclic AMP-response element Eva Rupp, Hubert Mayer and Edgar Wingender* Gesellschaft fOr Biotechnologische Forschung mbH, Department of Genetics, Mascheroder Weg 1, D-3300 Braunschweig, FRG Received July 11, 1990; Revised and Accepted Septenber 3, 1990

ABSTRACT We have screened the sequence of the 394 base pairs upstream of the main transcriptional start site of the promoter of the human parathyroid hormone (PTH) gene for well-known protein recognition motifs with the aim to identify potential positive or negative regulatory elements. Within this region we found a potential cAMP-response element (CRE) besides several other putative binding sites for transcription factors. We fused promoter regions that contain this element and extend beyond the transcription start site to an appropriate reporter gene (CAT) and transfected different cell lines with these constructs. Transient expression of the CAT gene from these hybrid genes could be shown to be significantly stimulated by forskolin or isoproterenol thus proving the responsiveness of the whole promoter region towards elevated cAMP levels. DNase I protection studies revealed protein binding around the putative CRE (PTHCRE) and an adjacent CCAAT element. Gel retardation assays with the PTH-CRE as well as the wellcharacterized CRE from the rat somatostatin promoter indicated specific binding of the same protein to both elements, although with a slightly reduced affinity of the PTH-CRE. Both of these elements were also able to confer cAMP-responsiveness to a heterologous promoter. INTRODUCTION Parathyroid hormone (PTH) is a peptide hormone that regulates the calcium homeostasis in the serum. It is synthesized and secreted by the parathyroid gland in response to low calcium levels, and this effect is in part due to alterations in the steady state level of PTH mRNA (1) and in part to posttranscriptional events (2). It has been demonstrated that the activity of protein kinase C in parathyroid cells is inversely correlated with extracellular calcium concentration (3). Recently, it has been shown that low extracellular calcium concentrations lead to an increase of PTH mRNA levels in the parathyroidea of rats (4). The second most important regulator for PTH synthesis is

*

To whom correspondence should be addressed

dihydroxycholecalciferol (1,25(OH)2D3) which decreases the PTH mRNA in the parathyroid cells in cell culture (5) or in the gland in vivo (6). Furthermore, there exist a number of reports concerning the influence of estradiol and progesterone on PTH secretion that might also refer to PTH transcription (7, 8). Moreover, secretion of PTH is stimulated either by cAMP directly (9, 10) or by ligands that have to act via surface membranes and intracellular second messenger pathways such as cyclo-AMP, e.g. ,B-adrenergic agonists (9, 11). However, no evidence has been accumulated until now that it is the transcription of the PTH gene which is (also) influenced by cAMP levels as it has been shown for a series of other genes (c-fos (12); fibronectin (13, 14); glycoprotein hormone a-subunit (15); PEPCK (13, 15); somatostatin (13, 16, 17)). The beststudied example probably is the rat somatostatin gene. In its promoter, a cAMP-response element (CRE) has been functionally defined, and this function has been ascribed to the canonical palindromic sequence 5'-TGACGTCA-3' (18). During the last two years, it became clear that a family of CRE-binding proteins (CREB) mediates its function (16, 19-22). They are closely related to the ATF proteins that bind to and stimulate the transcription of different adenovirus promoters and therefore all these factors are grouped into the ATF/CREB family (23, 24). At least one of them (CREB) seems to be stimulated in its transcription activating function after phosphorylation of its Ser-133 by the cAMP-dependent protein kinase A (25). It would be of great pharmacological and medical interest to elucidate the control of PTH synthesis through the cAMP-signal transduction pathway, e.g. by adrenergic effectors. Since ,Badrenergic receptors are known to be coupled to adenylate cyclase by a Gs protein, we attempted to characterize the cAMPresponsiveness of the PTH promoter and to identify its cAMPresponse element with regard to its protein binding and transcriptional enhancing properties.

MATERIALS AND METHODS Materials Plasmid pUC13

purchased from Boehringer Mannheim (Mannheim, FRG), pCAT3M (26) was a kind gift from P. Gruss was

5678 Nucleic Acids Research, Vol. 18, No. 19 (MPI, Gottingen, FRG) and pBLCAT2 (27) was kindly provided by G.Schutz (DKFZ, Heidelberg, FRG). For sequencing of double-stranded DNA a kit from Pharmacia was used. ROS 17/2.8 cells were used with kind permission of G.Rodan and T47D cells were purchased from ATCC. A plasmid expressing 3-galactosidase under control of the ,B-globin promoter (pHO3APr-I-(3Gal) was a gift of Martin Rentrop (DKFZ, Heidelberg). Dulbecco's modified Eagle medium was purchased from Flow laboratories, insulin, forskolin and isoproterenol from Sigma. Synthetic oligonucleotides were kindly provided by the DNA-synthesis group of the GBF. A fraction of HeLa cell extract which is enriched in CREB/ATF but also contains other DNAbinding factors was a kind gift by K.H. Seifart (University of Marburg, FRG); it has been proven that protein(s) contained in this fraction bind to the CRE of the c-fos promoter (K.H. Seifart, pers. communication). Poly[d(I:C)] used for gel retardation and footprinting experiments was purchased from Boehringer Mannheim and DNaseI grade I from Worthington (Freehold, New Jersey). All autoradiographies were done using a Kodak X-omat film.

Plasmid constructions After fill-in with Klenow DNA polymerase, either the TaqI/PstI (-254/+65) or the AflI/PstI (- 161/+65) fragment of the PTH promoter were cloned between the SmaI and the PstI sites of pUC 13. The resulting plasmids were termed pUC-TAQ or pUCAFL, respectively. CAT fusion plasmids were created by cloning the BglIl-fragment of the hPTH promoter that comprises the region between -698 and +61 into the BglII site of pCAT3M immediately upstream of the CAT-coding region. From the resulting plasmid (pCB41), the large PstI-fragment of 2.8 kb comprising the transcribed region of the PTH gene from position +2 to +65 and the chloramphenicol acetyltransferase (CAT) gene was re-cloned into the PstI site of either pUC-TAQ or pUCAFL giving rise to the test plasmids that were termed pP(-254)CAT and pP(- 161)-CAT, respectively. For DNaseI-footprinting assays, the RsaI/PstI-fragment from the human PTH promoter (28) was subcloned into pUC13 using the SmaI and PstI sites of this vector. The following pairwise complementary CRE oligonucleotides were synthesized: 5'-CGGAGTGACATCATCTGT-3' and 5'-CTAGACAGATGATGTCACTCCGCATG-3' representing the PTH-CRE; 5 '-CTGGCTGACGTCAGAGAT-3' and 5'-CTAGATCTCTGACGTCAGCCAGCATG-3' as rat somatostatin CRE (SOM-CRE) (16); 5'-CCAAGTGACTCAGCGCT-3' and 5'-CTAGAGCGCTGAGTCACTTGGCATG-3' represent the AP-1 binding site of human metallothionein HA (29, 30). Annealing of the complementary oligonucleotides provided a SphI and a XbaI site making them suitable for cloning between the corresponding sites of pBLCAT2. Positive clones were identified by dot-blot hybridization using one of the singlestranded oligonucleotides as probe, and were verified by dssequencing with T7-polymerase (31, 32). Cell culture and transfection ROS17/2.8 cells were grown in DMEM, 10% FCS and 4 mM glutamine, T47D cells in DMEM, 10% FCS, 4 mM glutamine and 0.6 /Ag/ml insulin. 24 hours before transfection, cells from a subconfluent culture were seeded into 94 mm diameter petridishes (3 x 105 cells/dish). 4 hours before transfection, the medium was removed, and 5 ml of fresh medium containing 2% FCS were added. For transfection of the cells of one dish, 5 Atg

DNA were coprecipitated with calcium phosphate as described by Graham and van der Eb (33). 2-5 isg of the plasmid pHi3APr-l-3Gal that governs expression of (-galactosidase were cotransfected for the determination of transfection efficiency. 18 hours after transfection, medium was replaced by medium containing either 1 x 10-6 M forskolin, 1 x 10-7 M isoproterenol or no additives. 24 hours after induction, cells were harvested and CAT as well as (3-galactosidase activities were determined as described (34-36). One induction experiment yielded 150 IL extract of which 90 IL were used for the CAT assay and 50 il for the determination of the ,B-galactosidase activity. Quantification of CAT activity was done by scanning the autoradiographies with a laser densitometer (LKB) and referring to the corresponding ,3-galactosidase activities.

Preparation and fractionation of nuclear extracts Nuclear extracts used for footprinting were prepared as described by Dignam et al. (37). Nuclear extracts designated for DNaseIfootprinting were fractionated using a 5 ml Heparin-Sepharose column, prepared according to the instructions of the manufacturer. Proteins eluted from the column applying steps of 0.1 M, 0.6 M and 1.0 M KCI were detected using the BioRad protein assay (38). Peak fractions were pooled and used for footprinting experiments. For gel retardation assays, extracts were prepared as described by Schreiber et al. (39). DNaseI protection experiments The RsaI/PstI PTH promoter fragment spanning position -393 to + 1, which was subcloned into PstI/SmaI sites of pUC 13, was excised with HindllI/EaeI, labelled with Klenow fragment of DNA polymerase I in the presence of [a32P]dATP (3000 Ci/mmol; Amersham Buchler, Braunschweig, FRG) and re-cut with EcoRI yielding a PTH promoter fragment labelled at the 3' end of the upper strand. Binding reactions were performed with 40,000 cpm of this labelled fragment in 10 mM HEPESNaOH, pH 7.9, 10 mM DTT, 7.5% PEG 6000, 8.5% glycerol, 0.2 mM EDTA and 1 yg poly[d(I:C)] for 15 min at room temperature. After incubation, 14 ug DNase I dissolved in 2 ILI of 10 mM HEPES-NaOH, pH 7.9, 0.2 mM EDTA, 1.25% glycerol, 7.5% PEG 6000, 125 mM MgCl2 were added and the samples were digested for 2 min at room temperature. Digestion was stopped by adding 45 yl of 75 mM EDTA, 0.75% SDS, 0.71 mg/ml proteinase K and incubating for 30 min at 37°. After subsequent phenol/ chloroform extraction and ethanol precipitation, samples were analyzed on a 8 % polyacrylamide/8 M urea sequencing gel. Gel retardation assay Reaction mixtures contained up to 24tg of protein, 24Lg poly[d(I:C)] and, if indicated, competing oligonucleotides in a final volume of 15 jl of 10 mM HEPES-NaOH, pH 7.9, 2.5 mM MgCl2, 10% glycerol, 50 mM KCI, 0.1 mM EDTA, 1 mM DTT. After preincubation at room temperature for 15 min, 130 fmol of oligonucleotide labelled with Klenow fragment of DNA-polymerase I were added and incubation continued for additional 15 min. Subsequently, 2.5 ILI of a stop solution containing 90 mM EDTA, 0.25% bromophenol blue and 10% sucrose were added to the samples which, after preelectrophoresis of the gels for 2 h, were loaded on a 6% polyacrylamide gel (30:0.5).

Nucleic Acids Research, Vol. 18, No. 19 5679

RESULTS Presence of a putative CRE in the human PTH promoter The nucleotide sequence of the human PTH promoter has been reported up to position -281 by Vasicek et al. (40) or up to position -805 by Reis et al. (28) if the first nucleotide of the human PTH gene that has been found in mature mRNA is assigned to the position + 1 (40). Since more precise mapping of the transcriptional start sites revealed some heterogeneity in each of the two reported clusters of transcripts (41) we use that previously defined site as position + 1 in this report. Using an updated version of the previously published data collection of transcription factor binding sites (42), we identified several potential recognition elements in the PTH promoter region. Among them are AP-1-homologs at -337 and -282, an AP-3 consensus sequence at -273, an octamer motif at -223, a putative SpI binding site at -144 with an adjacent CACCC motif in reverse orientation at - 135, and two CCAAT sequences in direct or in reverse orientation in position -175 or -89, resp. (Fig. 1). Most of these elements have been found to be protected against DNase I cleavage in the presence of nuclear cell extracts (E. Rupp et al., in preparation). A sequence resembling the canonical cAMP-response element 5'-TGACGTCA-3' is found at position -81 but deviates in one position in the center of the recognition motif (5'-TGACATCA-3'). This element is also present in the rat PTH promoter (43), and in the bovine PTH promoter this element matches even exactly the consensus sequence (44). Moreover, the bovine element has previously been shown to compete for protein binding to the CRE of the human DNA polymerase (3 gene (45) but has also been characterized -390

TG

-350

GTACGGTTATMCAMTACACTTATTTTTGGATTTTMTTTTCMGTMGTAAGATMTG ASTCA

TBASTCA

-300

ACTTAATCATAAACCTTTGAAATCAGTCTTTTTACAGTATMATCAGATTCATTGATTCA AP-1

AP-I

WWWCCACA ********

TGGCMNNTGCC *** ***** *

*

CMGCAT ****

TTAATCCACATAGAATTTTTCTCGATGGTATAATTCTGTATTTGTTMAAGTCTTTGCAT TaqI

AP-3

NF-I

OCTA

-200

CCAAT

AAbLLLLI ILMGbLLIbL1GA 1A l1lI TCG IATITAIGTATCC

AITITATAAAILM

AflIl -100 GACCMGACCC ** ******* *****. . ****** ** AAGAGTGTGCACCGCCCAATGGGTGTGTGTATGTGCTGCTTTGAACCTATAGTTGAGATC CAC COUP Spi ATTGG TGACGTCA ACCCTCTCTT CAT

as exhibiting only minimal cis-activity (46) or even as 'unfunctional' in comparison with rat somatostatin CRE (47).

cAMP responsiveness of the PTH core promoter To assay the PTH promoter for its responsiveness towards an elevated intracellular cAMP level, we fused the sequences between the sites for either TaqI at -254 or Afll at -161 and the PstI site at position +65 to the coding region of the prokaryotic chloramphenicol acetyltransferase (CAT) as a reporter gene using the pCAT3M vector (26). These constructs, pP(-254)-CAT or pP( - 161)-CAT, respectively, were transfected into cells of the clonal rat osteosarcoma cell line ROS 17/2.8 which we originally introduced in our experiments for investigation of vitamin D regulation mechanisms. CAT expression has been investigated either in the absence or presence of 10-6 M forskolin. This plant diterpene is known to stimulate adenylate cyclase (48) and we verified this for ROS 17/2.8 cells under the conditions used here as well (data not shown; Hollnagel et al., in preparation). Both constructs governed CAT expression at similar basal levels, and were identically and significantly stimulated by forskolin (Fig. 2). Quantification and correction for co-expressed (-galactosidase activities yielded a stimulation factor of 1.9i0.2 (p0.01; n=4). We also routinely elevated intracellular cAMP levels using isoproterenol, a /3-adrenergic agonist, whose receptor is coupled to adenylate cyclase. Applying this agent to cells transfected with the constructs mentioned above, we also reproducibly found an increased CAT activity (Fig. 2). Both constructs were significantly stimulated by 10-7 M isoproterenol (2.9-i0.7; p